PROBLEM 1 & 2
Here below are the solutions to Problem 1 and integrated plots from Problem 2.
A) What was the effect of time-of-day on mean RT and the standard deviation of RT?
Table 1: Reaction time mean and standard deviation with respect to time-of-day
| tod | mean | sd |
|---|---|---|
| 1 | 1249.723 | 840.1433 |
| 2 | 1267.193 | 723.7861 |
| 3 | 1121.494 | 599.8719 |
| 4 | 1171.105 | 1069.9249 |
In the afternoon people were more reactive and tended to vary less in reaction time compared to the other times of the day. In the morning and late-morning people were generally slower and showed great variation in reaction time. However, the evening time showed the most variation among people but people generally performed better.
B) What was the effect of session on mean RT?
After the 1st session participants were 23.19% faster in the second session and they fluctuated until the last 2 sessions where their reaction time dropped to 921ms and 958ms from the initial 1509ms performance,thus displaying a 587ms faster reaction time.
C)What was the effect of session on mean accuracy?
| session | mean_accuracy |
|---|---|
| 1 | 0.8122449 |
| 2 | 0.8031746 |
| 3 | 0.7841270 |
| 4 | 0.7400000 |
| 5 | 0.8035714 |
| 6 | 0.7952381 |
| 7 | 0.8961749 |
| 8 | 0.8190476 |
| 9 | 0.7523810 |
| 10 | 0.7942857 |
| 11 | 0.7285714 |
| 12 | 0.7809524 |
Accuracy faced an initial drop and quickly rised in the 7th session to only then drop again.
D)Was there improvement in RT/accuracy within the 1rst session? That is, for session 1, how did mean RT and mean accuracy change across trials? Note that you will need to filter to get just session 1.
This first plot represents a colormap of the mean of all reaction times values for each individual and each trial with respect to whether they were correct or failed during the trial. Notice that the red dot-line represents the mean of the subjects mean for each trial.
Here only the accuracy index of the mean data is displayed for each trial
Here, I thought that the previous accuracy displayed was not normalized to the reaction time and I reported this plot with may seem counterintuitive at first but correct in the end.
The first plot shows the effects of a variable (reaction time) on two independent variables (first and last trial) while the second one is a boxplot showing the reaction time in each different time of the day for the 4th trial.
E) What was the effect of time of day and the response variable on RT and accuracy?
| resp | tod | rt_mean | accuracy |
|---|---|---|---|
| Left Hand | 1 | 1307.453 | 0.7987421 |
| Right Hand | 1 | 1196.818 | 0.7809798 |
| Left Hand | 2 | 1260.313 | 0.8466454 |
| Right Hand | 2 | 1272.757 | 0.8242894 |
| Left Hand | 3 | 1153.660 | 0.8298611 |
| Right Hand | 3 | 1095.026 | 0.7800000 |
| Left Hand | 4 | 1188.294 | 0.7279412 |
| Right Hand | 4 | 1154.872 | 0.7569444 |
Here I made a barplot for the response variable (left or right) with respect to reaction time and displaying accuracy directly within the bars. It looks as if the most accuracy is reached in the late-morning while the opposite can be said in the evening. Nonetheless, the fastest reaction time happens to be in the afternoon. The left hand is the most accurate for all times but in the evening but it also tends to be slower. This is interesting because people are mostly right-handed while here it may not be the case.
Problem 3
##
## Attaching package: 'gplots'## The following object is masked from 'package:stats':
##
## lowess
Here are 4 plots generated from the function for subjects (3,2,7,8) and sessions (2,1,6,2) from left to right from top to bottom. Notice that I included an empty plot for subject 7 and session 6 because there is no such question in the data.